Namaste, I'm

Hardik Sharma

Data Scientist & Machine Learning Engineer | AI Researcher

I specialize in developing advanced AI solutions with expertise in Computer Vision, NLP, and Deep Learning. Currently focusing on research and development in the field of machine learning and data science.

Download Resume Contact Me

Experience

Dec 2024 - Present

Computer Vision Intern

Truens AI, Remote

Developed real-time deepfake detection system for banking KYC video calls, achieving 73.8% F1 score on private dataset and 70% aggregate F1 score across multiple evaluation scenarios. Engineered multimodal ID card verification system combining computer vision and NLP to detect both visual and textual forgeries in identification documents.

July 2024 - Present

Research Internship

AIISC, South Carolina, Remote

Implemented 11 advanced jailbreak techniques from the Trees of Attack paper EMNLP submission, evaluating vulnerability across 22 major LLMs including GPT-4, Claude, and LLaMA models. Discovered critical security gaps with 95.4% success rate on LLaMA-2 70B and 89.5% on GPT-4 using Persuasion Attack techniques.

May 2024 - Nov 2024

Data Science Intern

DeepAlgorithms Pvt. Ltd., Hyderabad, Remote

Achieved 88% authentication accuracy with industry-leading security metrics FAR 0.04%, FRR 0.08% for behavioral biometric system. Engineered spatial attention-based deep learning model that eliminated manual feature engineering, processing raw sensor data from accelerometer, gyroscope, and magnetometer inputs.

May 2024 - Present

Research Internship

Indian Institute of Technology, Mandi, HP

Engineered spatial attention-based deep learning model that outperformed previous approaches, achieving 90.4% precision (3% improvement) and 97.9% specificity (0.6% improvement) on complex odor prediction tasks. Implemented fragment-based transformer architecture for processing raw molecular data directly for more accurate classification.

May 2023 - July 2023

Summer Research Internship

National Institute of Technology, Hamirpur, HP

Mentored under Prof. Jyoti Srivastava of Computer Science Department and learned various NLP methods. Worked on BBC News Dataset Text Summarization using extractive summarization method. Summarized News text with a sample ROUGE1 score of 78% and ROUGE2 score of 74% using a Graph-based model.

Skills

Technical Expertise

I am a data science enthusiast with a strong foundation in machine learning, computer vision, and natural language processing. I have completed Bachelors degree in Data Science from IIT Madras, I have developed expertise in various AI technologies and frameworks.

Alongside my academic pursuits, I also enjoy engaging in short drives in my free time.

Computer Vision NLP Deep Learning DevOps TensorFlow PyTorch Scikit-learn Hugging Face NLTK spaCy OpenCV Keras Python SQL Java Docker AWS S3 Git Flask FastAPI
Hardik Sharma

Projects

Here are some of the projects I've worked on. Each represents different skills and challenges I've tackled in the field of data science and machine learning.

Facial Expression Recognition

Facial Expression Recognition

Computer Vision Deep Learning CNN

Built CNN using MobileNetV2 architecture with data balancing through oversampling minority emotion classes. Achieved 96.96% training accuracy and 84.34% validation accuracy using Adam optimizer with categorical crossentropy.

View Project
BDM Capstone Project

BDM Capstone Project

Data Analysis Business Intelligence Data Visualization

Analyzed 4-week sales trends with 20% increase from week 1 to week 3, followed by a 15% decrease in week 4. Found 80% of sales from 6 top-performing products, leading to inventory and promotion optimization. Calculated 15% profit increase per SKU, demonstrating revenue potential.

GitHub Watch Demo
Movie Sentiment Prediction

Movie Sentiment Prediction

Machine Learning Data Analysis NLP

Secured Rank of 97 out of 977 participants and achieved a F1 score of 0.832. Used Machine Learning for classification with a Stacking Classifier including Logistic Regression, Random Forest, GradientBoost, and AdaBoost with Quantile Data Transformation using 5-Fold Cross-Validation.

GitHub
Grocery Store

Grocery Store

Python Flask SQLAlchemy SQLite

Developed a web application using Flask serving REST APIs. Used Jinja2 templates for rendering HTML and employed SQLite for data management. Integrated Bootstrap for polished UI with pre-designed components. Implemented robust backend logic with RESTful API architecture.

GitHub

Education

April 2021 - May 2025

Bachelor of Science in Data Science and Applications

Indian Institute of Technology, Madras

Chennai, TN | CGPA: 8.0

Aug 2020 – July 2023

Bachelor of Computer Applications

Minerva College of Education

Delhi, India| CGPA: 8.36

April 2019 – March 2020

Senior Secondary 12th Non-Medical

Partap World School

Indora, HP| Percentage: 88.2%

April 2017 – March 2018

Higher Secondary 10th

Partap World School

Indora, HP | Percentage: 88.4%

Podcasts

Podcast Cover
Podcast with IITM BS Team (YouTube)
Podcast Cover
Podcast with Code Unveil Channel (30k+ views, YouTube)

Positions of Responsibility

Apr 2021 – Aug 2022

Group Lead

Indian Institute of Technology, Madras

Maintained communication with members, provided course updates, deadlines, and academic support, arranged meetings between group members, and participated in the decision-making of the house and the planning of events for the house.

Apr 2023 – Present

Editorial Wing Member, Adhyay

Indian Institute of Technology, Madras

Checking, proofreading, and sharing editorial content for approval as part of the Civil Services Society.

Apr 2019 – Mar 2020

Head Boy

Partap World School

Led student council, coordinated school events, and represented the student body in official functions.

Apr 2015 – Mar 2019

House Captain

Partap World School

Organized house activities, motivated team members, and led the house to success in inter-house competitions.

Achievements

Publications & Certifications

CIRSNet: A lightweight Chemometric Infrared Spectral Network for spectral analysis

Under 3rd review in IEEE Transactions in Instrumentations and Measurements

BIOFastNet: Biomedical Fragmented Attention Spectral Transformer Network for Infrared spectra analysis

Submitted in IEEE Transactions on Medical Imaging

Text Summarization Of Publicly Available BBC News Dataset On Kaggle

Global Journal For Research Analysis, Feb 2024

View Paper

Certifications

Workshops & Presentations

  • Generative AI at IHub IIT Mandi (August-October 2024)
  • Presented at research Graduate meeting in IIT Mandi (12-13 March 2024)
  • RailMadad and SushrutaAI selected for SIH internal hackathons

Get In Touch

Let's connect

Feel free to reach out to me for collaborations, research opportunities, or if you'd like to discuss anything related to data science, machine learning, or AI.

Email

hhardikssharmagmail.com

Phone

9877100914

Location

Indpur, Himachal Pradesh

Education

IIT Mandi, Himachal

Languages

English, Hindi, Punjabi